Norwegian University of Science and Technology Page 1 of 7

Department of Electronics and Telecommunications

Contact during exam:

Navn: Ulf Österberg
Tel: 46 83 61 43

Solutions, Final Exam in TFE4130 Wave Propagation

English

Saturday August 18 2012

Time: 09:00 - 13:00

Graded: September 8, 2012

Permitted Sources:
D: No written and handwritten examination support materials are permitted. A specified,
simple calculator is permitted.
 Page 2 of 7

Problem 1

a) A Hertzian dipole consists of a short conducting wire of length dl which is terminated in

two small disks/spheres, and carries a time-harmonic current I(t). The termination of
the current I(t) causes a pile-up of charges at the spheres Q+ and Q− (we let dl be the
vector from Q− to Q+ ). The deposited charge q follows the continuity equation, which
in integral form is
dq
I(t) = − . (1)
dt
Here, by convention, I(t) is the current from Q+ to Q− , and q is the net charge that is
deposited on Q+ . When I(t) varies harmonically with time, I(t) = I0 e−iωt , the solution
for q(t) is
q(t) = q0 e−iωt , (2)
where q0 = I/(iω).
This setup is identical to an oscillating dipole, or Hertzian dipole, with a dipole moment

p = q0 dle−iωt . (3)

For greatest simplicity, we assume vacuum everywhere around the dipole. The most
straightforward way to find the E&M field from a Hertzian dipole is via the solution for
the retarded potentials:

1 ρ(R)eikR
ˆ
V (r) = dv, (4)
4πǫ0 v R
µ0 J(R)eikR
ˆ
A(r) = dv. (5)
4π v R

In writing down the solution for V and A it is explicitly assumed that ρ(r, t) = ρ(r)e−iωt ,
J(r, t) = J(r)e−iωt and that we are working in the Lorenz gauge ∇·A(r, t) = −ǫµ∂t V (r, t).
Away from the source, the electric and magnetic fields are related by Ampere’s law in its
simplified form
i
E(r) = ∇ × H(r). (6)
ǫ0 ω
where again we assume E(r, t) = E(r)e−iωt with a similar expression for H.
Solving for A(r) we find

µ0 Idl ikr µIdl ikr

A(r) = e ẑ = e (r̂ cos θ − θ̂ sin θ). (7)
4πr 4πr
 Page 3 of 7

1
In spherical coordinates, H(r) = µ0
∇ × A(r) gives
 
Idl ∂ ikr
ikr ∂
H(r) = − e sin θ − e cos θ ϕ̂
4πr2 ∂r ∂θ
  (8)
Idl 2 1 1
= k sin θ − eikr ϕ̂.
4π ikr (ikr)2

From this, it is straightforward to find the E(r)-field:

 
i i 1 ∂ 1 ∂
E= ∇×H= (sin θHϕ )r̂ − (rHϕ ) θ̂ (9)
ǫ0 ω ǫ0 ω r sin θ ∂θ r ∂r

which gives
  
Idl 2 ikr 1 1
E(r) = − η0 k e 2 cos θ − r̂
4π (ikr)2 (ikr)3
   (10)
1 1 1
− sin θ − + θ̂ .
ikr (ikr)2 (ikr)3

b) See e.g. Fig. 11-3 in Cheng. In the near-field, the electric and magnetic fields are quasi-
static with the electric field polarized along both r̂ and θ̂. In the far field, the fields are
like plane waves with the electric field polarized along θ̂ and the magnetic field polarized
along ϕ̂.

c) The directive gain is the ratio of the radiation intensity in a solid angle (a two-dimensional
angle that an object subtends to at a point) to the average radiation intensity. The
directivity is the maximum directive gain of an antenna, i.e. the ratio of the maximum
radiated power to the average radiated power.

d) The radiation pattern is shown in Cheng p. 616.

e) In the far-field, the electric field of a Hertzian dipole is essentially

eikr0
E0 = Em F (θ, ϕ) , (11)
r0
where r0 is the distance from the antenna to the observation point. For two, antennas,
with separation distance d << r0 and d << r1 the total electric field is
 ikr0
eiξ+ikr1

e
E = E0 + E1 = Em F (θ, ϕ) + , (12)
r0 r1
 Page 4 of 7

where ξ is the phase difference between the current in the two antennas. Using r1 ≈
r0 − d sin θ cos ϕ in the far-field we may write
 
F (θ, ϕ) ikr0 iφ/2 Ψ
E = Em e e 2 cos , (13)
r0 2

where φ = ξ − kd sin θ cos ϕ, or that the magnitude is equal to

 
2Em  Ψ 
|E| = |F (θ, ϕ)| cos  . (14)
r0 2

Thus, the antenna pattern is described by the pattern function of the element factor and
the array factor.

Problem 2

k k
a) There was a typo in the exam where Eq. (1c) should read H1 − H2 = Js × n̂. The
procedure is detailed in Cheng and most other books on electrodynamics. For Maxwells
divergence equations, one may draw a small pillbox around the boundary between the
two media which by use of the divergence theorem, as the height of the pillbox becomes
small gives

D⊥ ⊥
1 − D2 = ρs , (15a)
B⊥
1 − B⊥
2 = 0. (15b)

For the curl-equations, one may use the curl-theorem on a rectangular loop extending
into both media. As the thickness of the loop approaches zero (i.e. the edges of the loops
approach the boundary), then the integration around the loop gives
k k
E1 − E2 = 0,
k k
H1 − H2 = Js × n̂.

b) TE-waves are waves where the electric field is polarized perpendicular to the propagation
direction, while TE-waves are waves where the magnetic field is polarized perpendicular
to the propagation direction. TEM cannot propagate unless there is a second conductor
inside the waveguide. That is, if there is no free current inside the waveguide, and
the displacement current lies in the transverse plane of the waveguide, there can be no
magnetic field in the conductor.
 Page 5 of 7

c) The equivalent circuit is shown in the figure below. Applying Kirchoff’s voltage law
around the circuit, and Kirchoff’s current law to the node directly after the inductance
gives the transmission line equations:

∂v(t, z) ∂i(t, z)
− = Ri(t, z) + L , (16a)
∂z ∂t
∂i(t, z) ∂v(t, z)
− = Gv(t, z) + C . (16b)
∂z ∂t

R∆z L∆z
i(z, t) i(z + ∆z, t)
+ +

v(z, t) G∆z C∆z v(z + ∆z, t)

− −
∆z

d) By cross-differentiating the transmission line equations, one may derive a second order
PDE for v(t, z)

∂ 2 v(t, z) ∂ 2 v(t, z) ∂v(t, z)

= LC + (RC + LG) + RGv(t, z). (17)
∂z 2 ∂t2 ∂t
This equation is linear in v(t, z) so we may decompose v(t, z) into Fourier-amplitudes
ˆ
′
v(t, z) = v(ω ′ , z)eiω t dω ′ . (18)

Inserting Eq. (18) into Eq. (17) gives

∂2
2
v(ω ′ , z) = γ 2 (ω ′ )v(ω ′ , z), (19)
∂z
where γ 2 (ω
p
′
) = (R + iω √
′
L)(G + iω ′ C). For the distortionless line, γ(ω ′ ) can be written
γ(ω ′ ) = R C/L + iω ′ LC. Thus, the solution for each Fourier-amplitude is
′
√ √ ′
V (ω ′ , z) = e−γ(ω )z V (ω ′ , 0) = e−R C/Lz V (ω ′ , 0)ei LCzω
√ ′
(20)
= e−R C/Lz V (ω ′ , 0)eiω z/c ,
 Page 6 of 7

√
where we have used that LC = 1c .
The solution is given by the inverse Fourier transform
1 −R√C/Lz
ˆ
′ ′
V (t, z) = e v(ω ′ , 0)eiω z/c−iω t dω ′
2π (21)
√
= e−R C/Lz V (t − z/c, 0).

Thus, the
p signal propagates with a velocity c and is attenuated exponentially with a
factor R C/L.

e) Since E = E(t − z/c)ŷ and B = B(t − z/c)x̂ the divergence equations are automatically
fulfilled. We also see that both E and B fulfill the wave equation since they are functions
of (t − z/c) only. It then suffices to check that ∇ × E = − ∂B
∂t
:

∇ × E = −∂z Ey x̂
(t − z/c)2
 
ωE0
=− exp − sin[ω(t − z/c)]x̂
c 2T 2 (22)
(t − z/c)2
 
E0 t − z/c
+ exp − cos[ω(t − z/c)]x̂.
c T2 2T 2

For ∂t B:
∂
− B = −∂t Bx x̂
∂t
(t − z/c)2
 
ωE0
=− exp − sin[ω(t − z/c)]x̂ (23)
c 2T 2
(t − z/c)2
 
E0 t − z/c
+ exp − cos[ω(t − z/c)]x̂.
c T2 2T 2

Thus, Maxwell’s equations are fulfilled inside the dielectric. The necessary boundary
conditions are automatically fulfilled for these field. E.g. there is no tangential com-
ponent of E and no perpendicular component of B, while the remaining two boundary
conditions allows one to find Js and ρs .
 Page 7 of 7

Problem 3

Question # Alt. a) Alt. b) Alt. c)

1 X
2 X
3 X
4 X
5 X
6 X
7 X
8 X